INDEX

Explanations

invented, fake, thing, parts, kinds

The neuron fires strongly on words that signal something is made up or hypothetical—e.g. “invent,” “invented,” “inventing,” “fake,” “illusion,” “scenario.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

’

-1.41

]$}

-1.40

″]

-1.38

еди

-1.36

虽然

-1.36

}*/

-1.36

 THEIR

-1.35

 Только

-1.35

除了

-1.34

chrift

-1.34

POSITIVE LOGITS



1.34

強力

1.24

pejo

1.23

頗

1.23

</h1>

1.21

 богат

1.20

保湿

1.19

</h2>

1.19

Ehr

1.18

取决于

1.17

Activations Density 10.701%